Project Summary One of the hallmark features of our brains is their enormous capacity to learn and to adapt to changing environments. Theories of memory storage in neural circuits largely focus on activity-dependent changes in synaptic weights ? e.g. long-term potentiation (LTP) ? as plausible learning correlates. But does our synapto- centric view of the cellular events underlying learning really capture the essence of memory engrams? Recent studies on the formation of engrams, or ?mnemic traces? (a concept introduced by Richard Semon in 1904), suggest that the ultimate step in memory engram participation is the suprathreshold activation of neurons. We thus propose a complementary, neurocentric view, in which the participants in a functionally active engram are at least partly determined by cell-autonomous regulation of the intrinsic excitability of individual neurons. In this view, synaptic plasticity controls the formation of reciprocal connectivity patterns within and between engrams, and thus remains an important factor in circuit plasticity and learning. Recent reports indicate that a substantial percentage of pyramidal cells do not engage (remain silent) or are extremely unreliable when these neural circuits are activated, even in primary sensory cortices upon presentation of appropriate stimuli. Here, we will make use of this phenomenon to provide a proof-of-principle demonstration of a role of changes in membrane excitability (?intrinsic plasticity?) in engram formation. We will test the hypothesis that intrinsic plasticity activates previously silent (?dormant?) or unreliable neurons and integrates them into reliable engrams, thus providing a mechanism to dynamically regulate engram composition. We propose that activation of dormant or unreliable neurons constitutes a memory trace in cortical circuits (?intrinsic theory? of memory), by enhancing the capacity for input pattern representation, by increasing the engram activation probability, or by promoting engram stability. This hypothesis will be tested using whole-cell patch-clamp recordings from L2/3 pyramidal cells in the primary somatosensory cortex (S1; barrel cortex) of awake mice, which will be paired with two-photon imaging of GCaMP6s-encoded population activity. We plan to enhance intrinsic excitability by two methods: a) repeated injection of depolarizing currents through the patch pipette (non-synaptic activation; test for the intrinsic nature of this type of plasticity when combined with blockade of synaptic transmission; note that ?intrinsic? refers to the expression phase, but that under physiological conditions synaptic activity will be needed for induction), or b) deflection of select groups of whiskers at active whisking frequency (10-20Hz). We will not only monitor intrinsic excitability, but will test whether neurons become responsive to whisker stimulation (activation of dormant neurons). To assess neuronal integration into engrams, we will image activity patterns in populations of 100-200 neurons. Finally, we will examine whether cholinergic signaling ? through downregulation of SK2- type K+ channels ? facilitates intrinsic plasticity. The R21 mechanism is appropriate, because we are at an early stage of exploring and developing critical tests for the intrinsic hypothesis of learning presented here.